aristanetworks / sonic

Open source drivers and initialization library for Arista platforms running SONiC
GNU General Public License v2.0
25 stars 30 forks source link

[chassis] sai switch create fails on Arista-7800R3A-36DM2-C72 #44

Closed arlakshm closed 2 years ago

arlakshm commented 2 years ago

The linecard Arista-7800R3A-36DM2-C72 does not come up with the sonic master image. The orchagent aborts because the sai_switch_create fails with the following errors

Jun  3 21:30:15.856526 str2-sonic-lc5-1 ERR syncd0#syncd: [06:00.0] SAI_API_SWITCH:platformInit:1465 Invalid instance id 06:00.0 received.
Jun  3 21:30:15.856526 str2-sonic-lc5-1 CRIT syncd0#syncd: [06:00.0] SAI_API_SWITCH:sai_driver_init:643 Error initializing driver, rc = -1.
Jun  3 21:30:15.856605 str2-sonic-lc5-1 CRIT syncd0#syncd: [06:00.0] SAI_API_SWITCH:brcm_sai_create_switch:1174 initializing SDK failed with error Operation failed (0xfffffff5).
Jun  3 21:30:15.856655 str2-sonic-lc5-1 ERR syncd0#syncd: :- sendApiResponse: api SAI_COMMON_API_CREATE failed in syncd mode: SAI_STATUS_FAILURE

The asic is detected, lscpi output show the asic pcied address is valid

admin@str2-sonic-lc5-1:~$ lspci | grep -i broadcom
06:00.0 Ethernet controller: Broadcom Inc. and subsidiaries Device 8852 (rev 02)
07:00.0 Ethernet controller: Broadcom Inc. and subsidiaries Device 8852 (rev 02)
10:00.0 Ethernet controller: Broadcom Inc. and subsidiaries NetXtreme BCM57762 Gigabit Ethernet PCIe (rev 01)
admin@str2-sonic-lc5-1:~$

asic_id is correct in device_metadata

admin@str2-7804-lc5-1:~$ sonic-cfggen -d --var-json DEVICE_METADATA
{
    "localhost": {
        "asic_id": "06:00.0",
        "asic_name": "Asic0",
        "bgp_asn": "65100",
        "buffer_model": "traditional",
        "cloudtype": "None",
        "default_bgp_status": "up",
        "default_pfcwd_status": "disable",
        "deployment_id": "1",
        "docker_routing_config_mode": "separated",
        "hostname": "str2-sonic-lc5-1",
        "hwsku": "Arista-7800R3A-36DM2-C72",
        "mac": "2c:dd:e9:6c:cc:7d",
        "max_cores": "16",
        "platform": "x86_64-arista_7800r3a_36dm2_lc",
        "region": "None",
        "switch_id": "5",
        "switch_type": "voq",
        "synchronous_mode": "enable",
        "type": "SpineRouter"
    }
}
admin@str2-7804-lc5-1:~$

sonic version

admin@str2-7804-lc5-1:~$ show vers

SONiC Software Version: SONiC.master.104554-b3e33d4f4
Distribution: Debian 11.3
Kernel: 5.10.0-12-2-amd64
Build commit: b3e33d4f4
Build date: Mon May 30 19:49:17 UTC 2022
Built by: AzDevOps@sonic-build-workers-001K5Z

please check if this is a platform issue or not. let me know if you need more info.

arlakshm commented 2 years ago

syslog_sairedis_logs.txt

Staphylo commented 2 years ago

Thanks for bringing this to our attention. I'll share your issue with the team and report back.

arlakshm commented 2 years ago

changed the switch_id from 5 to 6, seeing the same crash. syslog_2.txt

kenneth-arista commented 2 years ago

@arlakshm please close if this is no longer an issue.

arlakshm commented 2 years ago

Issue not seen anymore, closing this issue.